Machine Learning for Predicting Personality Traits from Eye Tracking

Recently, personality prediction holds significant importance in human centered systems, particularly in decision support systems, the smart industry, and the development of human machine interfaces. Eyes can reveal deep insights into a person's personality, so people’s visual behavior could better reflect their personality. In this research, we demonstrate that machine learning techniques can predict individual personality traits from eye movements and minimize biases and errors reported by self-reported questionnaires. We performed an experiment using Random Forest, Gradient Boosting, and Extreme Gradient Boosting algorithms to predict individuals' Big Five personality traits based on their visual behavior. The findings of this study demonstrate that the proposed methodology outperforms existing methods in terms of accuracy and reliability.